st.functions(..., pure=True) very noisy

[andreareina]

The generated function output its printed at every invocation, when once per distinct input is sufficient with pure functions. Settings can be used to quiet it down but then a lot of desirable output (including h.note()s) is also lost.

minimal example:

import hypothesis as h
import hypothesis.strategies as st

keyfuncs = (
    # choose a return type
    st.sampled_from(
        [
            st.one_of(st.integers(), st.floats(allow_nan=False)),
            st.integers(1, 5).flatmap(
                lambda n: st.tuples(
                    *[st.one_of(st.integers(), st.floats(allow_nan=False))] * n
                )
            ),
        ]
    )
    # get a function with that return type
    .flatmap(lambda strat: st.functions(like=lambda x: x, returns=strat, pure=True))
)


@h.given(f=keyfuncs)
def test_functions_noisy(f):
    h.note(f)
    f(None)
    f(None)
    assert False  # so we get the output

output:

$ pytest hypothesis-functions-noisy.py 
=========================================================================================== test session starts ===========================================================================================
platform linux -- Python 3.8.11, pytest-6.2.5, py-1.10.0, pluggy-0.13.1
rootdir: [snip], configfile: pyproject.toml
plugins: anyio-3.3.4, subtests-0.7.0, cov-2.12.1, hypothesis-6.39.3
collected 1 item                                                                                                                                                                                          

hypothesis-functions-noisy.py F

[snip]

----------------------------------------------------------------------------------------------- Hypothesis ------------------------------------------------------------------------------------------------
Falsifying example: test_functions_noisy(
    f=<function hypothesis-functions-noisy.<lambda>.<locals>.<lambda>>,
)
Called function: lambda x: <unknown>(x=None) -> 0
Called function: lambda x: <unknown>(x=None) -> 0